URUANA DE MINAS — 2011 MODEL-FREE SAME-LINE SELF-CONTROL AUDIT

Frozen city center: -16.0593, -46.2585
Frozen target: observed portion of r<=8 km in Project 3087
Direct variable: raw measured scalar total field MAGBRU
Primary spatial direction: derivative toward geographic north

NEW ADVERSARIAL SELF-CONTROL
Each of the same 14 flight lines is compared with itself farther north.
This removes between-line zero offsets and does not require subtracting IGRF to establish a sign reversal.

Frozen target:
- negative raw northward slope: 14/14 lines
- median raw slope: -1.554762 nT/km

Same-line 16–20 km north control:
- positive raw northward slope: 14/14 lines
- median raw slope: 0.962614 nT/km

Paired change (north control minus target):
- positive in 14/14 lines
- median change: 2.577798 nT/km
- minimum change across all lines: 0.549080 nT/km

IGRF reference:
- target median IGRF slope: +0.992857 nT/km
- north-control median IGRF slope: +1.022984 nT/km
Thus the reference main-field gradient remains positive in both zones while the RAW MEASURED FIELD changes sign.

FLIGHT-DIRECTION ADVERSARY
The 14 lines split exactly 7 northbound and 7 southbound in acquisition order.
Northbound: target all negative; north control all positive; medians -1.414286 -> 0.889667 nT/km.
Southbound: target all negative; north control all positive; medians -1.976190 -> 1.035560 nT/km.

This is important because a common temporal drift projected onto space would reverse its apparent northward sign when the aircraft heading reverses. The observed spatial sign does not: it is negative in the target for both headings and positive farther north for both headings.

DATE REPLICATION
Target lines span three acquisition dates:
- 2011-05-27
- 2011-05-28
- 2011-05-30
The target-negative / far-north-positive relation is retained across all three dates.

CONTROL-BAND ROBUSTNESS
A post-discovery adversarial family tested starts 10–30 km north and widths 4–12 km.
From a 14-km start onward, every band of width >=6 km gives positive raw slopes on all 14 lines.
From a 16-km start onward, every tested width 4–12 km gives positive raw slopes on all 14 lines.
Therefore the result is not dependent on the single 16–20 km control window.

CLAIM BOUNDARY
DEMONSTRATED by this audit:
A meridional spatial-derivative sign reversal exists directly in the raw measured scalar total magnetic field on the same 2011 flight lines: negative in the frozen Uruana target cap and positive farther north.

This result is model-free with respect to the existence of the raw sign reversal. IGRF is used only as an external reference showing that the expected main-field gradient remains positive in both zones.

NOT demonstrated by this audit:
A sign reversal of the vector Z component itself. That remains reserved for raw fX/fY/fZ or another local vector observation.

INFERENTIAL STATUS
The same-line control and broad control-band family are post-discovery adversarial checks, not preregistered confirmatory p-values. Their value is physical/model-artifact exclusion and direct demonstration of the observed derivative turnover.